Financial reinforcement learning library. Full pipeline for training and evaluating trading agents.
FinRL is a financial reinforcement learning library providing a complete pipeline to train and evaluate trading agents using deep reinforcement learning algorithms. It supports stock trading, crypto trading (via Binance/CCXT), and portfolio allocation. Features include multiple DRL algorithms (PPO, A2C, DDPG, etc.), backtesting engine, paper trading mode, and integration with major market data providers. Used in academic research and quantitative finance education.